GJWTracker β€” User Guide ======================= A web reimplementation of GJTRACKER by Lance Collister, W7GJ, whose original program has served EME operators since 1973. Everything runs in your browser: nothing is uploaded, and no account is needed. All times are UTC. WHAT IT IS FOR -------------- Two stations can only work each other over the moon when both can see the moon at the same time. That shared period is the "common window", and it changes every day as the moon moves north and south of the celestial equator. GJWTracker finds those windows, and tells you how good the path will be while they last. It also tracks the sun and several radio sources, which is how you check that your array points where you think it does. THE SCREEN ---------- The layout follows the original program, section by section. ### Section A β€” Date range Start and end date for PLANNING mode. Any range up to 400 days. The status box on the right reports the result of the last callsign lookup. ### Section B β€” Stations Two rows: your own station (Home) and the station you want to work (DX Stn). Display Tick to include this station. With both ticked, only the times when BOTH can see the object are listed β€” the common window. Callsign Type a callsign and press Enter, or click the πŸ” button, to look it up in the station file. The locator and coordinates fill in automatically, and any notes appear underneath. A trailing "*" searches by prefix: "PA*" finds the first Dutch station in the file. Once a callsign is found, "[" and "]" step backwards and forwards through the file. Locator A Maidenhead locator, 4, 6 or 8 characters. Press Enter or click "=" to calculate the coordinates. It also works the other way: leave the locator empty, type a latitude and longitude, and press "=" to get the locator. Lat/Lon Degrees, minutes, seconds and direction. Filled in automatically from the locator, or type them directly for a precise position. Above Elevation limits for the output. The defaults, 0 and 90, list Below every time the object is above the horizon. Enter a NEGATIVE "Above" value if your horizon drops away β€” from a hilltop the moon is workable before it formally rises. Set "Below" to 20 or so if your array has no elevation rotator and can only work the moon when it sits low. ### Section C β€” Object and operation Object Moon, Sun, or a fixed radio source (Cassiopeia A, Cygnus A, Taurus A, Virgo A, Sagittarius A). Choose "Other…" to enter your own right ascension and declination. Two stations can only be compared for the Moon. A common window is a lunar idea, so the DX row is disabled for anything else. Azimuth Restricts the output to a range of azimuth. Useful when trees, range buildings or mountains block part of your view. The first box is the eastern limit, the second the western; 0 and 360 show everything. Click the label to switch it to a UTC range instead, and list only the hours you are actually awake. Times are HHMM: 1800 to 0600 wraps through midnight. Operation PLANNING produces a text report over the date range. REAL TIME shows live positions, updated every second. ### Section D β€” Settings Increment How often a position is printed, 1 to 60 minutes. Rcvr noise temp Your receive system noise temperature in kelvin. Used for the degradation figure. 500 K is a reasonable 6 m value; a good 144 MHz system is nearer 100 K. Band The band you are working. Sets path loss and sky temperature scaling. Max dB degradation Only list times when degradation is smaller than this. Set 2 to find only the good days. 0 disables the filter. IARU region Your region, which selects the activity windows. Units Kilometres or miles, for the distances in the report. Lines per page Repeats the header every N lines, for printing. 0 runs continuously, which is better on screen. Clock hours The difference between your computer clock and UTC. Leave minus UTC at 0 if your clock is correct, which it almost always is on a modern machine. ### Section E β€” Buttons RUN Start the calculation. STOP Stop real-time tracking. SET DEFAULTS Save every field as your startup values. RESTART Clear the screen and reload those saved values. HELP This guide. READING THE PLANNING OUTPUT --------------------------- Two views of the same numbers, switched with the toggle above the output: Normal view A table: sorted columns with their units, sticky headers, elevation in green when the object is workable and red when it is not, degradation colour coded. Each day is split into its separate passes β€” moonrise (β†— MR) and moonset (β†˜ MS) β€” for the home station, so a window that spans midnight does not read as one block with a hole in it. A legend under the output explains the window letters and the phase notes. Each azimuth carries an arrow drawn at the exact bearing, and each elevation a small angle diagram β€” a horizon line with a ray at the real tilt β€” so you can see where to point without reading the numbers. Hover either for the precise value. TXT The fixed-width plain text of the original program, column for column. This is the one that survives being pasted into an email to a sked partner, and it is what Copy and Download produce regardless of which view is showing. Your choice is remembered. Print uses whichever view is on screen. Numbers use a comma for decimals and a space for thousands, as the original program does: 368 947 km, 16,19', 4,4 dB. Over a range of three days or more the Normal view opens with **Best opportunities**: the six days worth planning around, pinned to the top of the screen so they stay reachable while you scroll. Click one to jump to it in the table; use Hide to collapse the panel. Each entry has its own address, ending `#best_opportunity_1`, `_2` and so on. Clicking one puts that address in the bar, so copying the link and sending it to a sked partner takes them straight to the same entry. Everything that changes what the report says travels in the link: both stations, the dates, the band, and β€” importantly β€” the **elevation limits for both stations**. Those decide which rows exist at all, so a link without them would show the recipient a different sked from the one you sent. The same goes for increment, receiver temperature, maximum degradation, region, units and the azimuth or UTC range. Links are kept short, because some chat clients truncate long ones: - short keys, `c` for callsign, `d` for the DX station, `s` and `e` for the dates, `b` for band - dates as YYMMDD - the grid is omitted when the station directory already knows the callsign β€” the recipient's copy looks it up - anything left at its default is omitted entirely - the anchor is `#b1` rather than `#best_opportunity_1` A typical two-station link is around 90 characters instead of 157. Older links in the long form still work and are rewritten to the short one when they load. Ranking on degradation alone would mislead you. Over a typical three-month range the very lowest-degradation days also have the shortest windows and the lowest peak elevation β€” a day offering 0.2 dB and four minutes of mutual visibility is not an opportunity. So the score weighs three things, each measured against the range you asked for: degradation (weighted most, since it is the one you cannot work around), how long the window lasts, and peak elevation. The header for each day carries: RANGE Distance to the moon. It varies by about 13% over a month, which is worth around 2 dB on the path. P +6 DAYS Days past perigee (P) or apogee (A). Perigee is the good end. 15.76'SD Semi-diameter: the apparent radius of the moon in arc minutes. DX: Great-circle distance between the two stations. JD: Julian date. The columns: UTC Time. NOTES "NM" near new moon, "FM" near full moon. Both raise the noise the sun contributes; neither prevents contacts, and 6 m EME works during new moon in particular. W Activity window. "A" Asian, "E" European, "N" North American. The convention W6PO popularised in the 1970s, anchored on the moon rather than on you, so every station sees the same letter at the same instant: the Asian window is the first two hours after moonrise in Europe, the European window runs from then until moonrise in Japan, and the North American window is the remainder. AZIMUTH ELEV Where to point, for each station. The first pair is the home station, the second the DX station. DEC The moon's declination β€” its angular distance north or south of the celestial equator. One column, because it belongs to the moon and not to either station. This is what makes common windows change from day to day. POL Spatial polarization offset between the two stations, in degrees. Pure geometry: the difference in apparent orientation of a linear antenna seen from two points on earth. It does NOT include Faraday rotation, which the ionosphere adds on top and which nobody can predict reliably. A large spatial offset means one of you may need to rotate polarization, or wait β€” but Faraday can just as easily cancel it as add to it. SKY Β°K Sky temperature behind the moon. The galactic plane is hot and the moon passes through it twice a month. See the accuracy note below. DEGRADATION How much worse the path is than on the best possible day, in dB. Printed as a positive number of dB lost, as the original does: 0 is the ideal, and bigger is worse. Two things go into it: - The extra path loss when the moon is further away than perigee, 40Β·log₁₀(d/perigee). Forty, not twenty: the wave spreads on the way out and again on the way back. Perigee to apogee is 2.3 dB, which on its own decides some skeds. - The noise penalty, how far the sky behind the moon raises your system temperature above the coldest sky the moon ever passes in front of. That reference matters β€” the moon stays within about 28.5Β° of the ecliptic and never visits the cold galactic poles, so measuring against the coldest sky that exists would inflate every figure. Checked against a full week of real GJTRACKER V3.0 output: 0.116 dB RMS over 294 rows, and 1.2 dB against the 1.3 dB printed in the 2008 manual, which was not used for the fit. Never negative: 0 already means "as good as it gets". The output is colour coded, following the original: green 0 to 2 dB, yellow 2 to 4, orange 4 to 6, red worse. READING THE REAL-TIME DISPLAY ----------------------------- Live azimuth and elevation for each displayed station, refreshed every second, plus: Doppler Shift of your own echo, in Hz, for the selected band. Negative as the moon recedes. Worth knowing before you call CQ. Moonrise/set For today, at that station. Illumination Percentage of the disc lit. Polarization Spatial offset, when two stations are shown. The map shades the part of the world that can see the object, marks the point where it is directly overhead, and shows your stations. Anyone inside the light region can work the moon at that moment. THE STATION FILE ---------------- The lookup is built from **CALL3.TXT**, the reference directory of 6 m EME stations maintained by W7GJ and shared with WSJT β€” the same file the original GJTRACKER reads from disk. Around 6,500 stations, with the revision date shown in the status box when the page loads. On top of that it adds every station known to the ON4KST EME List on this site, which catches operators too new or too local to have reached CALL3 yet. Both load in the background; if either is unreachable the other still works. Note that CALL3.TXT is not exhaustive by design: its own header records that DL and UA stations not active on 6 m EME were removed to keep the file under the size WSJT once required. If a lookup comes up empty, type the locator by hand. Your own records always win: an imported CALL3.TXT is never overwritten. To use your own, import a WSJT-format CALL3.TXT. It is kept in your browser's local storage and stays private to your machine. The format is one record per line: W7GJ,DN27ub,,,,EME,MT,Lance,50: 4x 9el 1.5KW Only the callsign and locator are required; anything after them is shown as notes when the callsign is looked up. ACCURACY -------- Positions follow Meeus, *Astronomical Algorithms*, with the topocentric parallax correction that matters near the horizon β€” without it the moon's elevation is wrong by up to a degree, and every horizon-grazing window with it. Verified against two reference runs of the original program: the worked examples printed in GJHELP.PDF (2008), and a full GJTRACKER V3.0 run over seven days (2026). Azimuth and elevation agree to within 0.05Β°, polarization offset to about 1Β°, the lunar range to under 50 km, and the perigee/apogee day count exactly on all ten sampled days. Good through 2100. One column differs. **DEC** here is the moon's geocentric declination, which matches the 2008 reference exactly (to 0.02Β°). The V3.0 reference prints a value about 1.3Β° lower, scaling with the declination itself rather than with the observer's position, which is not a parallax signature and which none of geocentric, home-topocentric or DX-topocentric reproduces. Rather than fit a correction to a single run, this keeps the value that is verifiably correct. DEC is a property of the moon and is not used in any other calculation, so the difference affects nothing but that column. **Sky temperature is an approximation, and degradation inherits it.** The original program reads a real radio sky map from a data file, sampled in coarse cells β€” which is why its SKY column sits on the same value for many rows at a time. GJWTracker uses a smooth model of the galactic plane and halo instead, so its value changes on every row. Compared against a full week of real GJTRACKER V3.0 output, the sky runs between 0.6Γ— and 1.2Γ— the reference, median 0.9Γ—. The degradation formula itself is sound: feed the original's own sky values into it and it reproduces the original's degradation to 0.14 dB. All the disagreement you see is the sky model, not the physics on top of it. This matters for the absolute number, not for the shape: the good days are still the good days, and the ranking between two dates is reliable. Do not quote the figure as a measurement. If you can supply the original's GJSKYTEMP.DAT, it can be used directly and the approximation goes away. CREDIT ------ GJTRACKER was written by Lance Collister, W7GJ, and descends from TRACKER, first written in 1973. The original Windows program, its help file and its data files remain available at https://www.bigskyspaces.com/w7gj/ and are worth having. This web version is hosted by F5UKW and follows the original's layout and output format deliberately, so anyone who knows the program can use it without relearning it.